available under acc-by-nc-nd 4.0 international license · 2020. 11. 12. · 2 biofisika institute...

1 Structure of HIV-1 gp41 with its membrane anchors targeted by neutralizing antibodies 2

3 Christophe Caillat1,†, Delphine Guilligay1,†, Johana Torralba2, Nikolas Friedrich3, Jose L. Nieva2, 4

Alexandra Trkola3, Christophe Chipot4,5,6, François Dehez4,5 and Winfried Weissenhorn1* 5

6 1 Univ. Grenoble Alpes, CEA, CNRS, Institut de Biologie Structurale (IBS), 71, avenue des 7

Martyrs, 38000 Grenoble, France. 8 2 Biofisika Institute (CSIC, UPV/EHU) and Department of Biochemistry and Molecular Biology, 9

University of the Basque Country (UPV/EHU), 48080, Bilbao, Spain. 10 3 Institute of Medical Virology, University of Zurich, 8057 Zurich, Switzerland. 11 4 Laboratoire de Physique et Chimie Théoriques (LPCT), University of Lorraine, CNRS, 12

Boulevard des Aiguillettes, 54506 Vandoeuvre-lès-Nancy Cedex, France. 13 5 Laboratoire International Associé, CNRS and University of Illinois at Urbana-Champaign, 54506 14

Vandoeuvre-lès-Nancy Cedex, France. 15 6 Department of Physics, University of Illinois at Urbana-Champaign, 1110 West Green Street, 16

Urbana, Illinois 61801, USA. 17

18

† These authors contributed equally 19

*Correspondence to: [email protected] 20

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

.CC-BY-NC-ND 4.0 International licenseavailable under a(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprintthis version posted November 12, 2020. ; https://doi.org/10.1101/2020.11.12.379396doi: bioRxiv preprint

https://doi.org/10.1101/2020.11.12.379396http://creativecommons.org/licenses/by-nc-nd/4.0/

2

Abstract 37 38 The HIV-1 gp120/gp41 trimer undergoes a series of conformational changes in order to catalyze 39

gp41-induced fusion of viral and cellular membranes. Here, we present the crystal structure of gp41 40

locked in a fusion intermediate state by an MPER-specific neutralizing antibody. The structure 41

illustrates the conformational plasticity of the six membrane anchors arranged asymmetrically with 42

the fusion peptides and the transmembrane regions pointing into different directions. Hinge regions 43

located adjacent to the fusion peptide and the transmembrane region facilitate the conformational 44

flexibility that allows high affinity binding of broadly neutralizing anti-MPER antibodies. 45

Molecular dynamics simulation of the MPER Ab-induced gp41 conformation reveals the transition 46

into the final post-fusion conformation with the central fusion peptides forming a hydrophobic core 47

with flanking transmembrane regions. This, thus, suggests that MPER-specific broadly neutralizing 48

antibodies can block final steps of refolding of the fusion peptide and the transmembrane region, 49

which is required for completing membrane fusion. 50

51 Introduction 52

53

Viral fusion proteins catalyze virus entry by fusing the viral membrane with cellular 54

membranes of the host cell, thereby establishing infection. The HIV-1 envelope glycoprotein (Env) 55

is a prototypic class I fusion protein that shares common pathways in membrane fusion with class 56

II and III viral membrane fusion proteins 1-4. HIV-1 Env is expressed as a gp160 precursor 57

glycoprotein that is cleaved into the fusion protein subunit gp41 and the receptor binding subunit 58

gp120 by host furin-like proteases. Gp41 anchors Env to the membrane and associates non-59

covalently with gp120, thereby forming a stable trimer of heterodimers, the metastable Env 60

prefusion conformation 5,6. Orchestration of a series of conformational changes transforms energy-61

rich prefusion Env into the low-energy, highly stable gp41 post-fusion conformation, which 62

provides the free energy to overcome the kinetic barriers associated with bringing two opposing 63

membranes into close enough contact to facilitate membrane fusion 2,3. 64

HIV-1 gp41 is composed of several functional segments that have been shown or suggested 65

to extensively refold upon fusion activation: the N-terminal fusion peptide (FP), a fusion peptide 66

proximal region (FPPR), the heptad repeat region 1 (HR1), a loop region followed by HR2, the 67

membrane proximal external region (MPER), the transmembrane region (TMR), and a cytoplasmic 68

domain. Structures of native Env trimers in complex with different broadly neutralizing antibodies 69

revealed the conformation of the gp41 ectodomain lacking MPER in the native prefusion 70

conformation 7-12. Env interaction with CD4 results in opening of the closed prefusion trimer 13,14, 71




3

which includes the displacement of gp120 variable regions 1 and 2 (V1-V2) at the apex of the trimer 72

but no changes in gp41 15. This is required for the formation of a stable ternary complex of Env-73

CD4 with the co-receptor 16-18. Co-receptor binding positions prefusion gp41 closer to the host-cell 74

membrane 5 and induces a cascade of conformational changes in gp41. First, the fusion peptide is 75

repositioned by ~70 Å 9 to interact with the target cell membrane, generating a 110 Å extended 76

fusion-intermediate conformation 19,20 that bridges the viral and the host cell membrane 21. 77

Subsequent refolding of HR2 onto HR1 leads to the formation of the six-helix bundle core structure 78 22-24, which pulls the viral membrane into close apposition to the host-cell membrane and, thus, sets 79

the stage for membrane fusion 22. 80

Membrane fusion generates a lipid intermediate hemifusion state, that is predicted to break 81

and evolve to fusion pore opening 25, which is regulated by six-helical bundle formation 26,27. 82

Furthermore residues within FPPR, FP, MPER and TM have been as well implicated in fusion 28-32 83

indicating that final steps in fusion are controlled by the conformational transitions of the membrane 84

anchors into the final post-fusion conformation. 85

Here, we set out to understand the conformational transitions of the gp41 membrane 86

anchors. We show that the presence of the membrane anchors increases thermostability. However, 87

complex formation with a MPER-specific neutralizing nanobody induced an asymmetric 88

conformation of the membrane anchors, which constitutes a late fusion intermediate. We show that 89

this conformation can be targeted by MPER bnAbs consistent with the possibility that MPER-90

specific nAbs can interfere all along the fusion process until a late stage. Starting from the 91

asymmetric conformation, we used MD simulation based modelling to generate the final post-92

fusion conformation, which reveals a tight helical interaction of FP and TM in the membrane 93

consistent with its high thermostability. Our work, thus, elucidates the structural transitions of the 94

membrane anchors that are essential for membrane fusion, which can be blocked by MPER-specific 95

bnAbs up to a late stage in fusion. 96

97

Results 98

99

Gp41FP-TM interaction with 2H10. 100

Two gp41 constructs, one containing residues 512 to 581 comprising FP, FPPR and HR1 (N-101

terminal chain, chain N) and one coding for resides 629 to 715 including HR2, MPER and TM (C-102

terminal chain, chain C)( Fig. S1A) were expressed separately, purified and assembled into the 103

monodisperse trimeric complex gp41FP-TM (Fig. S1B). Gp41FP-TM reveals a thermostability of 104

> 95°C as measured by circular dichroism (Fig. S2A) indicating that the presence of FP and TMR 105




4

increases the thermostability by > 7°C compared to gp41 lacking FP and TM 33. In order to facilitate 106

crystallization, gp41FP-TM was complexed with the llama nanobody 2H10 34 in β-OG buffer and 107

purified by size exclusion chromatography (SEC)(Fig. S1C). To determine the stoichiometry of 108

binding, we performed isothermal titration calorimetry (ITC), which indicated that gp41FP-TM and 109

2H10 form a 3:1 complex with a KD of 2.1 +/- 0.9 µM (Fig. S2B). Interaction of gp41FP-TM with 110

2H10 was further confirmed by biolayer interferometry (BLI) analysis (Fig. S2C). 111

112

Crystal structure of gp41 in complex with 2H10 113

The structure of gp41FP-TM in complex with 2H10 was solved by molecular replacement to a 114

resolution of 3.8 Å (Table S1). The asymmetric unit contained trimeric gp41FP-TM bound to one 115

2H10 nanobody as indicated by ITC (Fig. S2B). The six-helix bundle structure composed of three 116

N-terminal and three C-terminal chains is conserved from HR1 residue A541 to HR2 residue L661 117

in all three protomers, and identical to previous structures 22,23. However, TMR and FP do not follow 118

the three-fold symmetry and their chains point into opposite directions (Fig. 1A). 2H10 interacts 119

with chain C-A (Fig. 1A and B) and induces a partially extended MPER conformation, including 120

a kink at L669 that positions the rest of MPER and TM (N674 to V693) at a 45° angle with respect 121

to the six-helix threefold symmetry axis. The corresponding N-terminal chain A (chain N-A) has 122

its FP disordered and FPPR from G527 to A533 is flexible, while the remaining FPPR and HR1 123

form a continuous helix (Fig. 1C). The chain C-A 2H10 epitope spans from residues Q658 to N671, 124

which is involved in a series of polar contacts with 2H10. These include interactions of gp41FP-125

TM E662 to 2H10 Y37, S668 and the carbonyl of D664 to R56, K665 to E95, N671 and the 126

carbonyl of A667 to R54, K655 to R97 and R93 contacts E95 to position it for interaction with 127

K665 (Fig. 1B). Notably, mutations of R56, R93, E95 and R97 have been shown to affect 128

interaction 34. Chain N-B of the second protomer forms a long continuous helix comprising FP, 129

FPPR and HR1 from residues L518 to D589 with the first six residues of FP being disordered. 130

Likewise, chain C-B folds into a continuous helix from M629 to A700 comprising HR2, MPER 131

and TM (Fig. 1D). Cα superimposition of chain C-B with MPER containing gp41 structures 33,35 132

yields root mean-square deviations of 0.55 Å and 0.29 Å (Fig. S3), indicating that the straight 133

helical conformation is the preferred conformation in threefold symmetrical gp41. In the third 134

protomer, chain N-C has a helical FP linked by flexible residues G531 to A533 to a short helix of 135

FPPR that bends at A541 with respect to helical HR1. Its corresponding chain C-C contains helical 136

HR2 and a flexible region from N671 to N674, which stabilizes a ~45° rotation of the remaining 137

MPER-TM helix that extends to residue R707 (Fig. 1E). Thus, the structure reveals flexible regions 138

within FPPR and MPER. FPPR flexibility is supported by strictly conserved G528 and G531, while 139




5

MPER has no conserved glycine residues. However, the same kink within L661 to F673 has been 140

observed in the MPER peptide structure 36, and in complex with bnAb 10E8 37. The N-terminal FP 141

residues 512 to 517 are disordered within the detergent micelle. Flexibility of this region in the 142

absence of membrane is consistent with NMR peptide structures that propose a flexible coil 143

structure of the N-terminal part of FP in solution followed by a helical region starting at L518 38 as 144

observed here. Based on the flexible linkage of FP and TMR, we propose that both FPPR and 145

MPER act as hinges during gp41 refolding leading to membrane fusion. 146

147

MD simulation of gp41FP-TM in a lipid bilayer 148

In order to test whether the structure is influenced by the presence of the detergent, we probed its 149

stability by MD simulation in a bilayer having the lipid composition of the HIV-1 envelope. This 150

confirmed that the structure is stable in a membrane environment during a 1 µs simulation as only 151

the flexibly linked FP of chain N-C moves within the bilayer during the simulation (Fig. S4A and 152

B). The tip of the 2H10 CDR3 dips into the bilayer (Fig. S4B), hence confirming the membrane-153

anchoring role of W100 for neutralization 34. 154

155

Neutralization activity of 2H10 depends on membrane interaction 156

The structure suggests that 2H10 induced the asymmetry within the membrane anchors. Crystal 157

packing effects on the conformation can be excluded, because only the C-terminus of the chain C-158

C helix is involved in crystal lattice contacts (Fig. S5). We therefore further evaluated 2H10 as a 159

neutralizing nanobody, which showed modest neutralization as a bi-head (bi-2H10), whereas 160

neutralization depended on W100 located at the tip of CDR3 34, a hall mark of MPER-specific 161

bnAbs 39. In order to improve the breadth and potency of monovalent 2H10, we increased its 162

membrane interaction capacity by changing CDR3 S100d to F (2H10-F) alone and in combination 163

with additional basic residues S27R, S30K and S74R (2H10-RKRF) within the putative 2H10 164

membrane-binding interface suggested by MD simulation (Fig. S4C). Wild type 2H10 did not show 165

significant neutralization against a panel of 10 clade B pseudo-viruses as shown previously 34, with 166

the exception of some weak neutralization of NL4-3 and SF163P3. However, both 2H10-F and 167

2H10-RKRF show improved potency and breadth neutralizing six and eight pseudo-viruses, 168

respectively, albeit with less potency than wild-type bi-2H10 and bnAb 2F5, the latter recognizing 169

an overlapping epitope (Table 1). This result, thus, confirms monovalent 2H10 as a modest anti-170

MPER Ab that neutralizes by engaging MPER and the membrane. 171

172

2H10 blocks fusion before the stage of lipid mixing 173




6

The efficacy of bi-2H10 and 2H10-RKRF for blocking membrane merging was further assessed in 174

peptide-induced lipid-mixing assays, whereas a vesicle population is primed for fusion by addition 175

of the N-MPER peptide containing the 2H10 epitope, which produces a fluorescence intensity spark 176

at time 20 s (Fig. 2A) 40. Under these experimental conditions, incorporation of the peptide into the 177

vesicles takes less than 10 sec 40. After 120 sec, the mixture was supplemented with target vesicles 178

fluorescently labeled with N-NBD-PE/N-Rh-PE (indicated by the arrow in Fig. 2A). The increase 179

in NBD intensity as a function of time followed the mixing of the target vesicle lipids with those of 180

the unlabeled vesicles (kinetic trace labeled ‘+N-MPER’). The increase in NBD fluorescence was 181

not observed when labeled target vesicles were injected in a cuvette containing unlabeled vesicles 182

not primed with peptide (‘no peptide’ trace). Lipid mixing was strongly attenuated when the 183

vesicles primed for fusion with N-MPER were incubated with bi-2H10 before addition of the target 184

vesicles (Fig. 2A, +N-MPER/+bi-2H10, dotted trace). Thus, the N-MPER-induced membrane 185

perturbations, which can induce fusion with target membranes, were inhibited by incubation with 186

bi-2H10. Comparison of the kinetics of the lipid-mixing blocking effect of 2H10-RKRF, bi-2H10 187

and the 2F5 Fab showed that the three antibodies inhibited both the initial rates and final extents of 188

lipid mixing induced by N-MPER (Fig. 2B). Using a control MPER peptide lacking the 2H10 and 189

2F5 epitopes for vesicle priming no inhibition of lipid mixing by 2H10-RKRF, bi-2H10 and 2F5 190

Fab was observed (Fig. 2C), indicating that the inhibitory effects depend on epitope recognition. 191

Fusion inhibition levels estimated as a function of the antibody concentration further confirmed the 192

apparent higher potency exhibited by the bi-2H10 (Fig. 2D). Lower concentrations of bi-2H10 193

compared to 2H10-RKRF were required to attain full blocking of the lipid-mixing process when 194

measured 20 sec (initial rates) or 300 sec (final extents) after target-vesicle injection (Fig. 2D and 195

E). The higher inhibitory potency of bi-2H10 indicates an avidity effect, which was also evident 196

when the concentration of the epitope-binding fragments was plotted (Fig. 2D and E, empty 197

squares and dotted line). Moreover, bi-2H10 appeared to block efficiently the process even at 198

2H10:N-MPER ratios below 1:3 (mol:mol), consistent with the involvement of peptide oligomers 199

in the promotion of membrane fusion. Based on these data, we suggest that both 2H10-RKRF and 200

bi-2H10 neutralize HIV-1 at the stage of lipid mixing. 201

202

GP41FP-TM interaction with MPER bnAbs 203

Although the 2H10 epitope overlaps with the 2F5 MPER epitope 41, the 2F5-bound peptide 204

structure 41, cannot be superimposed without major clashes with adjacent gp41 protomers. In 205

contrast, Cα superposition of the structures of 10E8 and LN01 in complex with MPER peptides 206

demonstrated possible binding to gp41FP-TM chain C-C (Fig. 3A and B). Furthermore, HCDR3 207




7

of both 10E8 and LN01 could make additional hydrophobic contacts with adjacent FP in this 208

binding mode. To confirm 10E8 and LN01 interaction, we performed immunoprecipitation of 209

gp41FP-TM with both bnAbs, which confirmed their interaction in vitro (Fig. S2D). We next 210

validated binding by bio-layer interferometry (BLI) using gp41FP-TM as analyte. This revealed 211

KDs of 0,2 nM for 10E8 and 34 nM for LN01 (Fig. 3C and D). We conclude that bnAbs 10E8 and 212

LN01 interact with gp41FP-TM with high affinity likely by inducing and stabilizing an asymmetric 213

gp41 conformation similar to the one observed in complex with 2H10 as suggested by the structural 214

modeling (Fig. 3 and B). 215

216

Building a post fusion conformation by MD simulation 217

In order to follow the final refolding of the membrane anchors we modeled the post fusion 218

conformation employing MD simulation. Assuming that the final post-fusion conformation shows 219

a straight symmetric rod-like structure we constructed a model of gp41 from the protomer 220

composed of the straight helical chains N-B and C-B (Fig. S6A and B). This conformation is also 221

present in the symmetric six-helix bundle structures containing either MPER 35 or FPPR and MPER 222 33 (Fig. S3). In this model, FP and TM do not interact tightly (Fig. S6B), which, however, does not 223

explain the increased thermostability induced by FP and TM (Fig. S2A). 1-µs MD simulation of 224

this model (Fig. S6B) in solution, rearranges the membrane anchors such that they adopt a compact 225

structure with trimeric FP interacting with adjacent TMs. Furthermore, the TMs kink at the 226

conserved Gly positions 690 and 691, as observed previously 42 (Fig. S6C). In order to recapitulate 227

the stability of the model in the membrane, we performed an additional 1-µs MD simulation of the 228

model (Fig. S6C) in a bilayer resembling the HIV-1 lipid composition, which relaxed the TM to its 229

straight conformation (Fig. 4A). The final structural model reveals tight packing of trimeric FP 230

flexibly linked to HR1 by FPPR G525 to G527 (Fig. 4B). HR2-MPER and TMR form continuous 231

helices with the TMRs packing against trimeric FP (Fig. 4A and C), which spans one monolayer 232

(Fig. 4A). As conserved tryptophan residues within MPER have been previously implicated in 233

fusion 28,30, we analyzed their structural role in the post fusion model. This reveals that the indole 234

ring of W666 is sandwiched between Leu669 and T536 and packs against L537. W670 makes a 235

coiled-coil interaction with S534, while W672 is partially exposed and packs against L669 and 236

T676. W678 binds into a hydrophobic pocket defined by I675, L679, I682 and adjacent FP/FPPR 237

residues F522 and A526. W680 is partially exposed, but reaches into a pocket created by the flexible 238

FPPR coil (Fig. S7). Thus, most of the tryptophan residues have structural roles in the post-fusion 239

conformation, hence providing an explanation for their functional role in fusion 28. The MPER 240

epitopes recognized by 10E8 and LN01 are exposed in the post-fusion model, but antibody docking 241




8

to this conformation produced major clashes, consistent with no expected binding to the final post 242

fusion conformation. 243

244

Structural transitions of gp41 245

A number of Env SOSIP structures revealed the native conformation of gp41 (Fig. 5A and 246

C) 9,43,44. The gp41FP-TM crystal structure and the model of its post fusion conformation provide 247

further insight into the path of conformational changes that native gp41 must undergo to adopt its 248

final lowest energy state conformation. The first major conformational changes in gp41 that take 249

place upon receptor binding are extension of HR1, FPPR and FP into a triple stranded coiled coil 250

with flexible linkers between FPPR and FP that projects FP ~115 Å away from its starting position 251

(Fig. 5D). Notably, such an early intermediate fusion conformation structure has been reported for 252

influenza hemagglutinin (HA) 45. This is likely followed by an extension and rearrangement of HR2 253

and MPER producing 11-15-nm long intermediates that connect the viral and cellular membranes 254 20,46. Gp41 refolding into the six-helix bundle structure then produces flexibly linked asymmetric 255

conformations of FPPR-FP and MPER-TMR anchored in the cellular and viral membranes, 256

respectively, as indicated by the gp41FP-TM structure. This intermediate conformation may bring 257

viral and cellular membranes into close proximity (Fig. 5E) or may act at the subsequent stage of 258

hemifusion (Fig. 5F). Further refolding and interaction of FP-FPPR and MPER-TM will generate 259

the stable post fusion conformation (Fig. 5G), a process that completes membrane fusion. 260

261

Discussion 262

Membrane fusion is an essential step of infection for enveloped viruses such as HIV-1, and requires 263

extensive conformational rearrangements of the Env prefusion conformation 7-9 into the final 264

inactive post-fusion conformation 2,3. The fusion model predicts that six-helix bundle formation 265

apposes viral and cellular membranes with FP and TM inserted asymmetrically in the cellular 266

membrane and the viral membrane, respectively 21. Here, we show that gp41 containing its 267

membrane anchors can adopt this predicted conformation, which is facilitated by flexible hinges 268

present in FPPR and MPER, thus corroborating their essential roles in membrane fusion 3,5,47. The 269

asymmetric conformation of the membrane anchors suggest further that bundle formation occurs 270

before pore formation as suggested previously 26,27. The membrane-fusion model proposes further 271

that final steps in fusion involves rearrangement and interaction of TM and FP 21, which is 272

confirmed by the MD-simulation model of the post-fusion conformation. Furthermore, the length 273

of the rod-like post-fusion structure of 13 nm lacking the C-C loop is consistent with the gp41 274

structure lacking FP and TM 48. 275




9

FP is helical in the gp41FP-TM-2H10 complex and the MD-based post-fusion 276

conformation, in agreement with NMR-based helical FP peptide models 49,50, although β-strand 277

structures of FP have been implicated in fusion as well 51. In comparison, in native Env 278

conformations, FP adopts multiple dynamic conformations that are recognized by broadly 279

neutralizing antibodies 43,44,52,53. In the post-fusion conformation, FP spans one monolayer of the 280

membrane, in contrast to suggested amphipathic helix-like interaction of FP with the outer layer of 281

the target cell membrane 54,55. 282

Furthermore, the coiled-coil interactions within FP and with TM in the post fusion model 283

explain the increased thermostability of gp41FP-TM compared to gp41 lacking FP and TM 33. We 284

propose that refolding of FP and TM can liberate additional free energy to catalyze final steps of 285

fusion. Hence, replacement of TM by a phosphatidylinositol (PI) anchor inhibits membrane fusion 286 56,57, akin to the GPI-anchored HA inhibition of influenza virus membrane fusion at the stage of 287

hemifusion 58. 288

Mutations in MPER and FPPR interfere with fusion 28,30,59,60, and mutations in TM block 289

fusion 31 or reduce fusion efficiency 32. Our structural model of the post-fusion conformation 290

predicts that most of these mutations affect the final post-fusion conformation, in agreement with 291

proposed interactions of FPPR and MPER, as well as FP and TM 61,62, thereby corroborating their 292

essential roles at late stages of membrane fusion. 293

Gp41FP-TM interaction with the 2H10 MPER-specific nanobody induces the asymmetric 294

conformation of the membrane anchors. In order to confirm that 2H10 is, indeed, a neutralizing 295

MPER-specific nanobody, we engineered increased 2H10 membrane binding, which improved 296

breadth and potency of 2H10 neutralization, in agreement with enlarged potency by increasing 297

membrane-binding of 10E8 63-65. This result, thus confirmed 2H10 as a modest anti-MPER 298

neutralizing antibody that recognizes both its linear epitope and membrane 34. Consistent with its 299

neutralization capacity, 2H10 inhibits membrane fusion at the stage of lipid mixing like 2F5 and 300

other anti-MPER bnAbs 40,66. Moreover, gp41FP-TM interacts with MPER bnAbs 10E8 37 and 301

LN01 42 in agreement with docking both structures onto the kinked chain C-C MPER epitope. 302

Notably, the kink in the MPER peptide in complex with 10E8 37 is similar to the chain C-C kink 303

and present in MPER peptide NMR structures 36,50. Furthermore, Ala mutations in the kink (671-304

674) affect cell-cell fusion and lower virus infectivity 67 corroborating the physiological relevance 305

of the kinked conformation. We therefore propose that 10E8 and LN01 binding to gp41FP-TM 306

induces similar asymmetry, as observed in the gp41FP-TM-2H10 structure by sampling the 307

dynamic states of the membrane anchors. 308




10

Our data, thus, indicate that MPER antibodies can act all along the gp41 refolding pathway 309

from blocking initial conformations of close to native Env 68-70 up to a late fusion intermediate state 310

that has already pulled viral and cellular membranes into close apposition. This thus, opens a long 311

temporal window of action for MPER bnAbs consistent with the findings that the half-life of 312

neutralization of MPER bnAbs is up to 30 minutes post virus exposure to target cells 71,72. 313

Furthermore, only one Ab per trimer may suffice to block final refolding of the membrane anchors 314

required for fusion. Finally, the presence of dynamic linkers connecting the core of viral fusion 315

proteins with their membrane anchors FP and TM must be a general feature of all viral membrane 316

fusion proteins. 317

318

Materials and Methods 319

320 Cell Lines 321

TZM-bl cells were obtained from NIH-AIDS Research and Reference Reagent Program (ARRRP) 322

and used for neutralization assays. TZM-bl cells were maintained in Dulbecco’s modified Eagle’s 323

medium supplemented with 10% fetal bovine serum, 100 units of Penicillin and 0.1 mg/ml of 324

Streptomycin while TZM-bl expressing the FcγRI cells were maintained in Dulbecco’s modified 325

Eagle’s medium supplemented with 10% fetal bovine serum, 0.025M Hepes, 50 µg/ml of 326

Gentamicin, 20 µg/ml of Blasticidin. 327

328 HIV-1 Primary Viruses 329

Env-pseudotyped viruses were prepared by co-transfection of HEK 293-T cells with plasmids 330

encoding the respective env genes and the luciferase reporter HIV vector pNLluc-AM as described 331 73. A full list of Env pseudotyped viruses generated with corresponding gene bank entry, subtype 332

and Tier information is provided in Table S2. 333

334 GP41 expression and purification 335

DNA fragments encoding HIV-1 Env glycoprotein amino acids 512 to 581 (N-terminal chain, chain 336

N) and residues 629 to 715 (C-terminal chain, chain C) were cloned into vectors pETM20 and 337

pETM11 (PEPcore facility-EMBL), respectively. Both constructs contain an N-terminal Flag-tag 338

(DDDDK sequence) and chain C contains additional two C-terminal arginine residues (Fig. S1A). 339

Proteins were expressed separately in E. coli strain C41(DE3)(Lucigen). Bacteria were grown at 340

37°C to an OD600nm of 0,9. Cultures were induced with 1mM IPTG at 37°C for 3h for gp41 chain 341

N and at 25°C for 20h for gp41 chain C. Cells were lysed by sonication in buffer A containing 20 342




11

mM Tris pH 8, 100 mM NaCl and 1% CHAPS (3-[(3-cholamidopropyl) diméthylammonio]-1-343

propanesulfonate (Euromedex). The supernatant was cleared by centrifugation at 53 000 g for 30 344

min. Gp41 chain N supernatant was loaded on a Ni2-sepharose column, washed successively with 345

Buffer A containing 1M NaCl and 1M KCl, then Buffer A containing 50 mM imidazole. Gp41 346

chain N was eluted in Buffer A containing 500 mM imidazole. Gp41 chain C was purified 347

employing the same protocol as for gp41 chain N. Gp41 chain N was subsequently cleaved with 348

TEV (Tobacco Etch Virus) protease for 2h at 20°C and then overnight at 4°C. After buffer exchange 349

with a mono Q column using buffer B (Buffer A with 0,5 M NaCl), uncleaved material and cleaved 350

His-tags were removed by a second Ni2+-sepharose column in buffer A. TEV-cleaved gp41 chain 351

C and chain N were then mixed in a molar ratio 4:1 and incubated overnight. To remove the excess 352

of gp41 chain C, the gp41 complex was loaded on a 3rd Ni2+-sepharose column in buffer A, washed 353

with buffer A containing 50 mM imidazole and eluted with buffer A containing 500 mM imidazole. 354

Subsequently the gp41 chain N TrxA-His-tag was removed by TEV digestion for 2h at 20°C and 355

overnight at 4°C. After buffer exchange with a mono Q column in buffer B uncleaved complex and 356

the TrxA-His-tag fusion were removed by a 4th Ni2+-sepharose column. The final gp41FP-TM 357

complex was concentrated and loaded onto a Superdex 200 size exclusion column (SEC) in buffer 358

C containing 20 mM Tris pH 8,0, 100 mM NaCl and 1% n-octyl β-D-glucopyranoside (Anatrace). 359

360

Nanobody 2H10 expression 361

2H10 encoding DNA was cloned into the vector pAX51 34 and expressed in the E. coli BL21(DE3) 362

strain (Invitrogen). Bacteria were grown at 37°C to an OD600nm of 0,7 and induced with 1mM IPTG 363

at 20°C for 20h. After harvesting by centrifugation, bacteria were resuspended in lysis buffer 364

containing 20 mM Hepes pH 7,5 and 100 mM NaCl. Bacteria were lysed by sonication and 365

centrifuged at 48 000g for 30 min. Cleared supernatant was loaded on Protein A sepharose column, 366

washed with lysis buffer and eluted with 0,1 M glycine pH 2,9. Eluted fractions were immediately 367

mixed with 1/5 volume of 1M Tris pH 9,0. 2H10 was then further purified by SEC on a superdex 368

75 column in PBS buffer. Mutants of 2H10, 2H10-F (S100d) and 2H10-RKRF (S27R, S30K, S74R 369

and S100d) were synthesized (Biomatik) and purified as described for the wild type. The 2H10 bi-370

head was purified as described 34. 371

372

Circular dichroism 373

CD measurements were performed using a JASCO Spectropolarimeter equipped with a 374

thermoelectric temperature controller. Spectra of gp41-TM were recorded at 20 ºC in 1 nm steps 375

from 190 to 260 nm in a buffer containing PBS supplemented with 1% n-octyl β-D-376




12

glucopyranoside. For thermal denaturation experiments, the ellipticity was recorded at 222 nm with 377

1ºC steps from 20º to 95ºC with an increment of 80ºC h-1, and an averaging time of 30 s/step. For 378

data analysis, raw ellipticity values recorded at 222 nm were converted to mean residue ellipticity. 379

380

Isothermal Titration Calorimetry (ITC) 381

The stoichiometry and binding constants of 2H10 binding to gp41 FP-TM was measured by ITC200 382

(MicroCal Inc.). All samples used in the ITC experiments were purified by SEC in a buffer 383

containing 20 mM Tris pH 8.0, 100 mM NaCl and 1 % n-octyl β-D glucopyranoside and used 384

without further concentration. Samples and were equilibrated at 25 °C before the start of the 385

experiment. The ITC measurements were performed at 25 °C by making 20 2-μl injections of 267 386

µM 2H10 to 0.2 ml of 19.5 µM gp41FP-TM. Curve fitting was performed with MicroCal Origin 387

software. Three experiments were performed, with an average stoichiometry N = 1.1 +/- 0.2 2H10 388

binds to gp41FP-TM with a KD of 2.1 µM +/- 0.9. 389

390

Bio-layer Interferometry Binding Analysis 391

Binding measurements between antibodies (10E8 IgG, LN01 IgG and 2H10) were carried out on 392

an Octet Red instrument (ForteBio). For the determination of the binding between antibodies and 393

gp41FP-TM, 10E8 IgG or LN01 IgG or 2H10 were labelled with biotin (EZ-Link NHS-PEG4-394

Biotin) and bound to Streptavidin (SA) biosensors (ForteBio). The biosensors loaded with the 395

antibodies were equilibrated in the kinetic buffer (20 mM Tris pH 8.0, 100 mM NaCl and 1 % n-396

octyl β-D glucopyranoside) for 200-500 sec prior to measuring association with different 397

concentrations of gp41FP- for 100-200 seconds at 25 °C. Data were analyzed using the ForteBio 398

analysis software version 11.1.0.25 (ForteBio). For 10E8 the kinetic parameters were calculated 399

using a global fit 1:1 model and 2:1 model. For the determination of the binding of LN01 IgG and 400

2H10, KDs were estimated by steady state analysis. All bio-layer interferometry experiments were 401

conducted at least three times. 402

403

Immunoprecipitation of gp41FP-TM by bnAbs 10E8 and LN01 404

220 µg of Gp41FP-TM were incubated alone or with 50µg of 10E8 or LN01 antibodies for 10 h at 405

20°C in buffer C. The complex was loaded on Protein A sepharose affinity resin and incubated for 406

1h. The resin was subsequently washed 3 times with buffer C and eluted with SDS gel loading 407

buffer and boiling at 95°C for 5 min. Samples were separated on a 15% SDS-PAGE and stained 408

with Coomassie brilliant blue. 409

410




13

Neutralization assay 411

The neutralization activity of the 2H10 variants and mAbs was evaluated using TZM-bl cells and 412

Env pseudotyped viruses as described 73. Briefly, serial dilutions of inhibitor were prepared in cell 413

culture medium (DMEM with 10% heat-inactivated FBS, 100 U/ml penicillin and 100 μg/ml 414

streptomycin (all from Gibco)) and added at a 1:1 volume ratio to pseudovirus suspension in 384 415

well plates (aiming for 500’000–5’000’000 relative light units (RLU) per well in the absence of 416

inhibitors). After one-hour incubation at 37°C, 30 µl of virus-inhibitor mixture was transferred to 417

TZM-bl cells in 384 well plates (6000 cells/well in 30µl cell culture medium supplemented with 418

20µg/ml DEAE-Dextran seeded the previous day). The plates were further incubated for 48 hours 419

at 37°C before readout of luciferase reporter gene expression on a Perkin Elmer EnVision 420

Multilabel Reader using the Bright-Glo Luciferase Assay System (Promega). 421

The inhibitor concentration (referring to the mix with cells, virus and inhibitor) causing 50% 422

reduction in luciferase signal with respect to a reference well without inhibitor (inhibitory 423

concentration IC50) was calculated by fitting a non-linear regression curve (variable slope) to data 424

from two independent experiments using Prism (GraphPad Software). If 50% inhibition was not 425

achieved at the highest inhibitor concentration tested, a greater than value was recorded. To control 426

for unspecific effects all inhibitors were tested for activity against MuLV envelope pseudotyped 427

virus. 428

429

Fusion assay. 430

The peptides used in the fusion inhibition experiments, NEQELLELDKWASLW 431

NWFNITNWLWYIK (N-MPER) and KKK-NWFDITNWLWYIKLFIMIVGGLV-KK (C-MPER), 432

were synthesized in C-terminal carboxamide form by solid-phase methods using Fmoc chemistry, 433

purified by reverse phase HPLC, and characterized by matrix-assisted time-of-flight (MALDI-434

TOF) mass spectrometry (purity >95%). Peptides were routinely dissolved in dimethylsulfoxide 435

(DMSO, spectroscopy grade) and their concentration determined by the Biscinchoninic Acid 436

microassay (Pierce, Rockford, IL, USA). 437

Large unilamellar vesicles (LUV) were prepared following the extrusion method of Hope et al. 74. 438

1-palmitoyl-2-oleoylphosphatidylcholine (POPC) and cholesterol (Chol) (Avanti Polar Lipids, 439

Birmingham, AL, USA) were mixed in chloroform at a 2:1 mol:mol ratio and dried under a N2 440

stream. Traces of organic solvent were removed by vacuum pumping. Subsequently, the dried lipid 441

films were dispersed in 5 mM Hepes and 100 mM NaCl (pH 7.4) buffer, and subjected to 10 freeze-442

thaw cycles prior to extrusion 10 times through 2 stacked polycarbonate membranes (Nuclepore, 443

Inc., Pleasanton, CA, USA). Lipid mixing with fusion-committed vesicles was monitored based on 444




14

the resonance energy transfer assay described by Struck et al. 75, with the modifications introduced 445

by Apellaniz et al. 40. The assay is based on the dilution of co-mixed N-(7-nitro-benz-2-oxa-1,3-446

diazol-4-yl)phosphatidylethanolamine (N-NBD-PE) and N-(lissamine Rhodamine B 447

sulfonyl)phosphatidylethanolamine (N-Rh-PE) (Molecular Probes, Eugene, OR, USA), whereby 448

dilution due to membrane mixing results in increased N-NBD-PE fluorescence. Vesicles containing 449

0.6 mol % of each probe (target vesicles) were added at 1:9 ratio to unlabeled vesicles (MPER 450

peptide-primed vesicles). The final lipid concentration in the mixture was 100 µM. The increase in 451

NBD emission upon mixing of target-labeled and primed-unlabeled lipid bilayers was monitored 452

at 530 nm with the excitation wavelength set at 465 nm. A cutoff filter at 515 nm was used between 453

the sample and the emission monochromator to avoid scattering interferences. The fluorescence 454

scale was calibrated such that the zero level corresponded to the initial residual fluorescence of the 455

labeled vesicles and the 100 % value to complete mixing of all the lipids in the system (i.e., the 456

fluorescence intensity of vesicles containing 0.06 mol % of each probe). Fusion inhibition was 457

performed with bi-2H10, 2H10-RKRF and 2F5 Fabs at concentrations of 10 µg/ml and 20 µg/ml 458

as indicated. 459

460

Crystallization, data collection and structure determination 461

For crystallization, 1 mg of gp41FP-TM was mixed with 1.5 mg of 2H10. The complex was purified 462

by SEC on a Superdex 200 column in a buffer containing 100 mM NaCl, 20 mM Tris pH 8,0 and 463

1% n-octyl β-D-glucopyranoside and concentrated to 7-10 mg/ml. Crystal screening was performed 464

at the EMBL High Throughput Crystallization Laboratory (HTX lab, Grenoble) in 96-well sitting 465

drop vapor diffusion plates (Greiner). Following manual refinement of crystallization conditions, 466

crystals of gp41FP-TM in complex with 2H10 were grown by mixing 1 µl of protein with 1 µl of 467

reservoir buffer containing 0,1 M sodium citrate pH 6,0, 0,2 M ammonium sulfate, 20% 468

polyethylene glycol 2000 and 0,1 M NaCl at 20°C (293 K) in hanging drop vapor diffusion plates. 469

Before data collection, crystals were flash frozen at 100K in reservoir solution supplemented with 470

1% n-octyl β-D-glucopyranoside and 25 % ethylene glycol for cryo-protection. 471

Data were collected on the ESRF beamline ID30b at a wavelength of 0.9730 Å. Data were 472

processed with the program XDS 76. The data from two crystals were merged with Aimless 77. The 473

data set displayed strong anisotropy in its diffraction limits and was submitted to the STARANISO 474

Web server 78. The merged STARANISO protocol produced a best-resolution limit of 3.28 Å and 475

a worst-resolution limit of 4.74 Å at a cutoff level of 1.2 for the local Imean / σ(Imean) (note that 476

STARANISO does not employ ellipsoidal truncations coincident with the crystal axes). The 477

gp41FP-TM-2H10 crystals belong to space group C 2 2 21 and the structure was solved by 478




15

molecular replacement using the program Phaser 79 and pdb entries 1env and 4b50. The model was 479

rebuilt using COOT 80 and refined using Phenix 81. Data up to 3.28 Å were initially used for model 480

building but were finally truncated to 3.8 Å. Statistics for data reduction and structure refinement 481

are presented in Table S1. 482

One copy of gp41FP-TM in complex with 2H10 are present in the asymmetric unit. Numbering of 483

the nanobody 2H10 was performed according to Kabat. The gp41FP-TM-2H10 complex was 484

refined to 3.8 Å data with an R/Rfree of 26.7 / 31.1 %. 99.6 % of the residues are within the most 485

favored and allowed regions of a Ramachandran plot 77. Some of the crystallographic software used 486

were compiled by SBGrid 82. Atomic coordinates and structure factors of the reported crystal 487

structures have been deposited in the Protein Data Bank (https://www.rcsb.org; PDB: 7AEJ. 488

489

Figure Generation 490

Molecular graphics figures were generated with PyMOL (W. Delano; The PyMOL Molecular 491

Graphics System, Version 1.8 Schrödinger, LLC, http://www.pymol.org). 492

493

Molecular Dynamics (MD) simulation 494

Molecular assays. Starting from the crystal structure determined herein, we built two molecular 495

assays based (i) on the structure of the entire Gp41FP-TM/2H10 complex, and (ii) based on a gp41 496

model generated by a three-fold symmetrization of the straight helical chains N-B and C-B. 497

Electron density for FP and TM is partially absent in the crystal structure and the missing parts have 498

been built as helical extensions; FP from residue 512 to 518 and TM from residues 700 to 709 based 499

on TM structures (6SNE and 6B3U). All residues were taken in their standard protonation state. 500

The first assay included a fully hydrated membrane composed of 190 cholesterol, 40 1-palmitoyl-501

2-oleoyl-glycero-3-phosphocholine (POPC), 88 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-502

ethanolamine (POPE), 36 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (POPS) and 56 N-503

stearoyl sphingomyelin, present in the HIV-1 lipid envelope 83, using the CHARMM-GUI interface 504 84,85. The resulting molecular assembly consisted of about 178,000 atoms in a rhomboidal cell of 505

106 x 106 x 169 Å³. The second computational assay featured a water bath of 91 x 91 x 114 Å³, 506

representing a total of nearly 95,700 atoms. Both assays were electrically neutral, with a NaCl 507

concentration set to 150 mM. 508

Molecular Dynamics. All simulations were performed using the NAMD 2.14 program 86. Proteins, 509

cholesterol, lipids and ions were described using the CHARMM forcefield 87-89 and the TIP3P 510

model 90 was used for water. MD trajectories were generated in the isobaric-isothermal ensemble 511

at a temperature of 300 K and a pressure of 1 atm. Pressure and temperature were kept constant 512



https://www.rcsb.org/http://www.pymol.org/https://doi.org/10.1101/2020.11.12.379396http://creativecommons.org/licenses/by-nc-nd/4.0/

16

using the Langevin thermostat and the Langevin piston method 91, respectively. Long-range 513

electrostatic interactions were evaluated by the particle-mesh Ewald (PME) algorithm 92. Hydrogen 514

mass repartitioning 93 was employed for all simulations, allowing for using a time step of 4 fs. 515

Integration was performed with a time step of 8 and 4 fs for long- and short-range interactions, 516

respectively, employing the r-RESPA multiple time-stepping algorithm 94. The SHAKE/RATTLE 517 95,96 was used to constrain covalent bonds involving hydrogen atoms to their experimental lengths, 518

and the SETTLE algorithm 97 was utilized for water. 519

The computational assay formed by gp41 in an aqueous environment was simulated for a period of 520

1 µs, following a thermalization of 40 ns. For the gp41FP-TM/2H10 complex, the lipid bilayer was 521

first thermalized during 200 ns using soft harmonic restraints on every dihedral angle of the protein 522

backbones, allowing the complex to align optimally with its membrane environment. Following the 523

equilibration step, a production run of 1 µs was performed. 524

The final structure of the hydrated gp41 was also embedded in the HIV-1-like envelope membrane 525

employed for the gp41FP-TM/2H10 complex. The same 200 ns equilibration protocol was applied 526

followed by a production run of 1 µs. 527

528 References 529 530 531 1. Kielian, M. & Rey, F.A. Virus membrane-fusion proteins: more than one way to make a 532

hairpin. Nat Rev Microbiol 4, 67-76 (2006). 533 2. Weissenhorn, W., Hinz, A. & Gaudin, Y. Virus membrane fusion. FEBS Lett 581, 2150-5 534

(2007). 535 3. Harrison, S.C. Viral membrane fusion. Nat Struct Mol Biol 15, 690-698 (2008). 536 4. Baquero, E., Albertini, A.A. & Gaudin, Y. Recent mechanistic and structural insights on 537

class III viral fusion glycoproteins. Curr Opin Struct Biol 33, 52-60 (2015). 538 5. Chen, B. Molecular Mechanism of HIV-1 Entry. Trends Microbiol (2019). 539 6. Wang, Q., Finzi, A. & Sodroski, J. The Conformational States of the HIV-1 Envelope 540

Glycoproteins. Trends Microbiol 28, 655-667 (2020). 541 7. Julien, J.P. et al. Crystal Structure of a Soluble Cleaved HIV-1 Envelope Trimer. Science 542

342, 1477-83 (2013). 543 8. Lyumkis, D. et al. Cryo-EM Structure of a Fully Glycosylated Soluble Cleaved HIV-1 544

Envelope Trimer. Science 342, 1484-90 (2013). 545 9. Pancera, M. et al. Structure and immune recognition of trimeric pre-fusion HIV-1 Env. 546

Nature 514, 455-61 (2014). 547 10. Kwon, Y.D. et al. Crystal structure, conformational fixation and entry-related interactions 548

of mature ligand-free HIV-1 Env. Nat Struct Mol Biol 22, 522-31 (2015). 549 11. Stewart-Jones, G.B. et al. Trimeric HIV-1-Env Structures Define Glycan Shields from 550

Clades A, B, and G. Cell 165, 813-26 (2016). 551 12. Ward, A.B. & Wilson, I.A. The HIV-1 envelope glycoprotein structure: nailing down a 552

moving target. Immunol Rev 275, 21-32 (2017). 553 13. Liu, J., Bartesaghi, A., Borgnia, M.J., Sapiro, G. & Subramaniam, S. Molecular 554

architecture of native HIV-1 gp120 trimers. Nature 455, 109-13 (2008). 555




17

14. Ozorowski, G. et al. Open and closed structures reveal allostery and pliability in the HIV-556 1 envelope spike. Nature 547, 360-363 (2017). 557

15. Wang, H., Barnes, C.O., Yang, Z., Nussenzweig, M.C. & Bjorkman, P.J. Partially Open 558 HIV-1 Envelope Structures Exhibit Conformational Changes Relevant for Coreceptor 559 Binding and Fusion. Cell Host Microbe 24, 579-592 e4 (2018). 560

16. Shaik, M.M. et al. Structural basis of coreceptor recognition by HIV-1 envelope spike. 561 Nature 565, 318-323 (2019). 562

17. Dobrowsky, T.M., Zhou, Y., Sun, S.X., Siliciano, R.F. & Wirtz, D. Monitoring early 563 fusion dynamics of human immunodeficiency virus type 1 at single-molecule resolution. J 564 Virol 82, 7022-33 (2008). 565

18. Chang, M.I., Panorchan, P., Dobrowsky, T.M., Tseng, Y. & Wirtz, D. Single-molecule 566 analysis of human immunodeficiency virus type 1 gp120-receptor interactions in living 567 cells. J Virol 79, 14748-55 (2005). 568

19. Frey, G. et al. A fusion-intermediate state of HIV-1 gp41 targeted by broadly neutralizing 569 antibodies. Proc Natl Acad Sci U S A 105, 3739-44 (2008). 570

20. Lai, R.P. et al. A fusion intermediate gp41 immunogen elicits neutralizing antibodies to 571 HIV-1. J Biol Chem 289, 29912-26 (2014). 572

21. Weissenhorn, W. et al. Structural basis for membrane fusion by enveloped viruses. Mol 573 Membr Biol 16, 3-9 (1999). 574

22. Weissenhorn, W., Dessen, A., Harrison, S.C., Skehel, J.J. & Wiley, D.C. Atomic structure 575 of the ectodomain from HIV-1 gp41. Nature 387, 426-30 (1997). 576

23. Chan, D.C., Fass, D., Berger, J.M. & Kim, P.S. Core structure of gp41 from the HIV 577 envelope glycoprotein. Cell 89, 263-73 (1997). 578

24. Caffrey, M. et al. Three-dimensional solution structure of the 44 kDa ectodomain of SIV 579 gp41. Embo J 17, 4572-84 (1998). 580

25. Chernomordik, L.V. & Kozlov, M.M. Membrane Hemifusion: Crossing a Chasm in Two 581 Leaps. Cell 123, 375-382 (2005). 582

26. Melikyan, G.B. et al. Evidence that the transition of HIV-1 gp41 into a six-helix bundle, 583 not the bundle configuration, induces membrane fusion. J Cell Biol 151, 413-23 (2000). 584

27. Markosyan, R.M., Cohen, F.S. & Melikyan, G.B. HIV-1 envelope proteins complete their 585 folding into six-helix bundles immediately after fusion pore formation. Mol Biol Cell 14, 586 926-38 (2003). 587

28. Salzwedel, K., West, J.T. & Hunter, E. A conserved tryptophan-rich motif in the 588 membrane-proximal region of the human immunodeficiency virus type 1 gp41 589 ectodomain is important for Env-mediated fusion and virus infectivity. J Virol 73, 2469-590 80 (1999). 591

29. Peisajovich, S.G., Epand, R.F., Pritsker, M., Shai, Y. & Epand, R.M. The polar region 592 consecutive to the HIV fusion peptide participates in membrane fusion. Biochemistry 39, 593 1826-33 (2000). 594

30. Bellamy-McIntyre, A.K. et al. Functional links between the fusion peptide-proximal polar 595 segment and membrane-proximal region of human immunodeficiency virus gp41 in 596 distinct phases of membrane fusion. J Biol Chem 282, 23104-16 (2007). 597

31. Shang, L., Yue, L. & Hunter, E. Role of the membrane-spanning domain of human 598 immunodeficiency virus type 1 envelope glycoprotein in cell-cell fusion and virus 599 infection. J Virol 82, 5417-28 (2008). 600

32. Long, Y., Meng, F., Kondo, N., Iwamoto, A. & Matsuda, Z. Conserved arginine residue in 601 the membrane-spanning domain of HIV-1 gp41 is required for efficient membrane fusion. 602 Protein Cell 2, 369-76 (2011). 603

33. Buzon, V. et al. Crystal structure of HIV-1 gp41 including both fusion peptide and 604 membrane proximal external regions. PLoS Pathog 6, e1000880 (2010). 605




18

34. Lutje Hulsik, D. et al. A gp41 MPER-specific llama VHH requires a hydrophobic CDR3 606 for neutralization but not for antigen recognition. PLoS Pathog 9, e1003202 (2013). 607

35. Shi, W. et al. Structural characterization of HIV gp41 with the membrane-proximal 608 external region. J Biol Chem 285, 24290-8 (2010). 609

36. Sun, Z.Y. et al. HIV-1 broadly neutralizing antibody extracts its epitope from a kinked 610 gp41 ectodomain region on the viral membrane. Immunity 28, 52-63 (2008). 611

37. Huang, J. et al. Broad and potent neutralization of HIV-1 by a gp41-specific human 612 antibody. Nature 491, 406-12 (2012). 613

38. Serrano, S. et al. Structure-Related Roles for the Conservation of the HIV-1 Fusion 614 Peptide Sequence Revealed by Nuclear Magnetic Resonance. Biochemistry 56, 5503-5511 615 (2017). 616

39. Cerutti, N., Loredo-Varela, J.L., Caillat, C. & Weissenhorn, W. Antigp41 membrane 617 proximal external region antibodies and the art of using the membrane for neutralization. 618 (2017). 619

40. Apellaniz, B. et al. Cholesterol-dependent membrane fusion induced by the gp41 620 membrane-proximal external region-transmembrane domain connection suggests a 621 mechanism for broad HIV-1 neutralization. J Virol 88, 13367-77 (2014). 622

41. Ofek, G. et al. Structure and mechanistic analysis of the anti-human immunodeficiency 623 virus type 1 antibody 2F5 in complex with its gp41 epitope. J Virol 78, 10724-37 (2004). 624

42. Pinto, D. et al. Structural Basis for Broad HIV-1 Neutralization by the MPER-Specific 625 Human Broadly Neutralizing Antibody LN01. Cell Host Microbe 26, 623-637 e8 (2019). 626

43. Yuan, M. et al. Conformational Plasticity in the HIV-1 Fusion Peptide Facilitates 627 Recognition by Broadly Neutralizing Antibodies. Cell Host Microbe 25, 873-883 e5 628 (2019). 629

44. Kumar, S. et al. Capturing the inherent structural dynamics of the HIV-1 envelope 630 glycoprotein fusion peptide. Nat Commun 10, 763 (2019). 631

45. Benton, D.J., Gamblin, S.J., Rosenthal, P.B. & Skehel, J.J. Structural transitions in 632 influenza haemagglutinin at membrane fusion pH. Nature 583, 150-153 (2020). 633

46. Ladinsky, M.S. et al. Electron tomography visualization of HIV-1 fusion with target cells 634 using fusion inhibitors to trap the pre-hairpin intermediate. Elife 9(2020). 635

47. Blumenthal, R., Durell, S. & Viard, M. HIV entry and envelope glycoprotein-mediated 636 fusion. J Biol Chem 287, 40841-9 (2012). 637

48. Weissenhorn, W. et al. The ectodomain of HIV-1 env subunit gp41 forms a soluble, alpha-638 helical, rod-like oligomer in the absence of gp120 and the N-terminal fusion peptide. 639 EMBO J 15, 1507-14 (1996). 640

49. Li, Y. & Tamm, L.K. Structure and Plasticity of the Human Immunodeficiency Virus 641 gp41 Fusion Domain in Lipid Micelles and Bilayers. Biophys J. 93, 876-885 (2007). 642

50. Serrano, S. et al. Structure and immunogenicity of a peptide vaccine, including the 643 complete HIV-1 gp41 2F5 epitope: implications for antibody recognition mechanism and 644 immunogen design. J Biol Chem 289, 6565-80 (2014). 645

51. Apellaniz, B., Huarte, N., Largo, E. & Nieva, J.L. The three lives of viral fusion peptides. 646 Chem Phys Lipids 181, 40-55 (2014). 647

52. Blattner, C. et al. Structural delineation of a quaternary, cleavage-dependent epitope at the 648 gp41-gp120 interface on intact HIV-1 Env trimers. Immunity 40, 669-80 (2014). 649

53. Kong, R. et al. Fusion peptide of HIV-1 as a site of vulnerability to neutralizing antibody. 650 Science 352, 828-33 (2016). 651

54. Han, X., Bushweller, J.H., Cafiso, D.S. & Tamm, L.K. Membrane structure and fusion-652 triggering conformational change of the fusion domain from influenza hemagglutinin. Nat 653 Struct Biol 8, 715-20 (2001). 654




19

55. Lai, A.L., Moorthy, A.E., Li, Y. & Tamm, L.K. Fusion activity of HIV gp41 fusion 655 domain is related to its secondary structure and depth of membrane insertion in a 656 cholesterol-dependent fashion. J Mol Biol 418, 3-15 (2012). 657

56. Weiss, C.D. & White, J.M. Characterization of stable Chinese hamster ovary cells 658 expressing wild-type, secreted, and glycosylphosphatidylinositol-anchored human 659 immunodeficiency virus type 1 envelope glycoprotein. J Virol 67, 7060-6 (1993). 660

57. Salzwedel, K., Johnston, P.B., Roberts, S.J., Dubay, J.W. & Hunter, E. Expression and 661 characterization of glycophospholipid-anchored human immunodeficiency virus type 1 662 envelope glycoproteins. J Virol 67, 5279-88. (1993). 663

58. Kemble, G.W., Danieli, T. & White, J.M. Lipid-anchored influenza hemagglutinin 664 promotes hemifusion, not complete fusion. Cell 76, 383-391 (1994). 665

59. Munoz-Barroso, I., Salzwedel, K., Hunter, E. & Blumenthal, R. Role of the membrane-666 proximal domain in the initial stages of human immunodeficiency virus type 1 envelope 667 glycoprotein-mediated membrane fusion. J Virol 73, 6089-92 (1999). 668

60. Narasimhulu, V.G.S. et al. Distinct functions for the membrane-proximal ectodomain 669 region (MPER) of HIV-1 gp41 in cell-free and cell-cell viral transmission and cell-cell 670 fusion. J Biol Chem 293, 6099-6120 (2018). 671

61. Louis, J.M. et al. Insights into the Conformation of the Membrane Proximal Regions 672 Critical to the Trimerization of the HIV-1 gp41 Ectodomain Bound to Dodecyl 673 Phosphocholine Micelles. PLoS One 11, e0160597 (2016). 674

62. Reuven, E.M. et al. HIV-1 gp41 transmembrane domain interacts with the fusion peptide: 675 implication in lipid mixing and inhibition of virus-cell fusion. Biochemistry 51, 2867-78 676 (2012). 677

63. Chen, J. et al. Mechanism of HIV-1 neutralization by antibodies targeting a membrane-678 proximal region of gp41. J Virol 88, 1249-58 (2014). 679

64. Kwon, Y.D. et al. Surface-Matrix Screening Identifies Semi-specific Interactions that 680 Improve Potency of a Near Pan-reactive HIV-1-Neutralizing Antibody. Cell Rep 22, 1798-681 1809 (2018). 682

65. Rujas, E. et al. Functional Optimization of Broadly Neutralizing Hiv-1 Antibody 10e8 by 683 Promoting Membrane Interactions. J Virol (2018). 684

66. Huarte, N. et al. The broadly neutralizing anti-human immunodeficiency virus type 1 685 4E10 monoclonal antibody is better adapted to membrane-bound epitope recognition and 686 blocking than 2F5. J Virol 82, 8986-96 (2008). 687

67. Sun, Z.Y. et al. Disruption of helix-capping residues 671 and 674 reveals a role in HIV-1 688 entry for a specialized hinge segment of the membrane proximal external region of gp41. J 689 Mol Biol 426, 1095-108 (2014). 690

68. Lee, J.H., Ozorowski, G. & Ward, A.B. Cryo-EM structure of a native, fully glycosylated, 691 cleaved HIV-1 envelope trimer. Science 351, 1043-8 (2016). 692

69. Rantalainen, K. et al. HIV-1 Envelope and MPER Antibody Structures in Lipid 693 Assemblies. Cell Rep 31, 107583 (2020). 694

70. Carravilla, P. et al. Molecular recognition of the native HIV-1 MPER revealed by STED 695 microscopy of single virions. Nat Commun 10, 78 (2019). 696

71. Shen, X. et al. Prolonged exposure of the HIV-1 gp41 membrane proximal region with 697 L669S substitution. Proc Natl Acad Sci U S A 107, 5972-7 (2010). 698

72. Williams, L.D. et al. Potent and broad HIV-neutralizing antibodies in memory B cells and 699 plasma. Sci Immunol 2(2017). 700

73. Rusert, P. et al. Divergent effects of cell environment on HIV entry inhibitor activity. 701 AIDS 23, 1319-27 (2009). 702




20

74. Hope, M.J., Bally, M.B., Webb, G. & Cullis, P.R. Production of large unilamellar vesicles 703 by a rapid extrusion procedure: characterization of size distribution, trapped volume and 704 ability to maintain a membrane potential. Biochim Biophys Acta 812, 55-65 (1985). 705

75. Struck, D.K., Hoekstra, D. & Pagano, R.E. Use of resonance energy transfer to monitor 706 membrane fusion. Biochemistry 20, 4093-9 (1981). 707

76. Kabsch, W. Integration, scaling, space-group assignment and post-refinement. Acta 708 Crystallogr D Biol Crystallogr 66, 133-44 (2010). 709

77. Evans, P.R. & Murshudov, G.N. How good are my data and what is the resolution? Acta 710 Crystallogr D Biol Crystallogr 69, 1204-14 (2013). 711

78. Tickle, I.J. et al. STARANISO. (http://staraniso.globalphasing.org/cgi-bin/staraniso.cgi). 712 Cambridge, United Kingdom: Global Phasing Ltd. (2018). 713

79. McCoy, A.J. et al. Phaser crystallographic software. J Appl Crystallogr 40, 658-674 714 (2007). 715

80. Emsley, P., Lohkamp, B., Scott, W.G. & Cowtan, K. Features and development of Coot. 716 Acta Crystallogr D Biol Crystallogr 66, 486-501 (2010). 717

81. Adams, P.D. et al. PHENIX: a comprehensive Python-based system for macromolecular 718 structure solution. Acta Crystallogr D Biol Crystallogr 66, 213-21 (2010). 719

82. Morin, A. et al. Collaboration gets the most out of software. Elife 2:e01456(2013). 720 83. Brugger, B. et al. The HIV lipidome: a raft with an unusual composition. Proc Natl Acad 721

Sci U S A 103, 2641-6 (2006). 722 84. Jo, S., Kim, T., Iyer, V.G. & Im, W. CHARMM-GUI: a web-based graphical user 723

interface for CHARMM. J Comput Chem 29, 1859-65 (2008). 724 85. Wu, E.L. et al. CHARMM-GUI Membrane Builder toward realistic biological membrane 725

simulations. J Comput Chem 35, 1997-2004 (2014). 726 86. Phillips, J.C. et al. Scalable molecular dynamics with NAMD. J Comput Chem 26, 1781-727

802 (2005). 728 87. MacKerell, A.D. et al. All-atom empirical potential for molecular modeling and dynamics 729

studies of proteins. J Phys Chem B 102, 3586-616 (1998). 730 88. Klauda, J.B. et al. Update of the CHARMM all-atom additive force field for lipids: 731

validation on six lipid types. J Phys Chem B 114, 7830-43 (2010). 732 89. Best, R.B. et al. Optimization of the additive CHARMM all-atom protein force field 733

targeting improved sampling of the backbone phi, psi and side-chain chi(1) and chi(2) 734 dihedral angles. J Chem Theory Comput 8, 3257-3273 (2012). 735

90. Jorgensen, W.L., Chandrasekhar, J., Madura, J.D., Impey, R.W. & Klein, M.L. 736 Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 737 926−935 (1983). 738

91. Feller, S.E., Zhang, Y.H., Pastor, R.W. & Brooks, B.R. Constant pressure molecular 739 dynamics simulations − The Langevin piston method. J. Chem. Phys. 1995, 103, 740 4613−4621. J. Chem. Phys. 103, 4613−4621 (1995). 741

92. Darden, T., York, D. & Pedersen, L. Particle mesh Ewald: An N•log (N) method for 742 Ewald sums in large systems. . J. Chem. Phys. 98, 10089−10092 (1993). 743

93. Hopkins, C.W., Le Grand, S., Walker, R.C. & Roitberg, A.E. Long-Time-Step Molecular 744 Dynamics through Hydrogen Mass Repartitioning. J Chem Theory Comput 11, 1864-74 745 (2015). 746

94. Tuckerman, M.E., Berne, B.J. & Martyna, G.J. Reversible multiple time scale molecular 747 dynamics. . J. Chem. Phys. 97, 1990−2001 (1992). 748

95. Andersen, H.C. Rattle: a “velocity” version of the shake algorithm for molecular dynamics 749 calculations. . J. Comput. Phys. 52, 24-34 (1983). 750



http://staraniso.globalphasing.org/cgi-bin/staraniso.cgihttps://doi.org/10.1101/2020.11.12.379396http://creativecommons.org/licenses/by-nc-nd/4.0/

21

96. Ryckaert, J.-P., Ciccotti, G. & Berendsen, H.J.C. Numerical integration of the cartesian 751 equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. 752 Comput. Phys. 23, 327–341. J Comput Phys 23, 327–341 (1977). 753

97. Miyamoto, S. & Kollman, P.A. SETTLE: An analytical version of the SHAKE and 754 RATTLE algorithms for rigid water models. . J. Comput. Chem. 13, 952−962 (1992). 755

98. Li, Z. et al. Subnanometer structures of HIV-1 envelope trimers on aldrithiol-2-inactivated 756 virus particles. Nat Struct Mol Biol 27, 726-734 (2020). 757

758 759 Acknowledgement 760

W.W. acknowledges support from the Institute Universitaire de France (IUF), from the European 761

Union's Horizon 2020 research and innovation programme under grant agreement No. 681137, 762

H2020 EAVI and the platforms of the Grenoble Instruct-ERIC center (IBS and ISBG; UMS 3518 763

CNRS-CEA-UGA-EMBL) within the Grenoble Partnership for Structural Biology (PSB). Platform 764

access was supported by FRISBI (ANR-10-INBS-05-02) and GRAL, a project of the University 765

Grenoble Alpes graduate school (Ecoles Universitaires de Recherche) CBH-EUR-GS (ANR-17-766

EURE-0003). J.L.N. acknowledges funding from Spanish MINECO (BIO2015-64421-R; 767

MINECO/AEI/FEDER, UE), Spanish MCIU (RTI2018-095624-B-C21; MCIU/AEI/FEDER, UE) 768

and Basque Government (IT1196-19). We thank Miriam Hock and Serafima Guseva for previous 769

contributions to the project, the ESRF-EMBL Joint Structural Biology Group for access and support 770

at the ESRF beam lines, J. Marquez (EMBL) from the HTX crystallization facility, C. Mas and J.-771

B. Reiser for assistance on ISBG platforms. 772

773

Author Contributions: W.W. conceived and designed the study. J.L.N. supervised the lipid 774

mixing experiments performed by J.T. A.T. supervised neutralization assays performed by N.F. 775

C.C. and F.D. performed molecular dynamics simulation experiments. D.G. produced proteins, 776

mutants and performed crystallization and pull-down experiments. C.Ca. performed the structural 777

studies and interaction studies. W.W. wrote the manuscript with input from all authors. 778

779

Competing interests: The authors declare no competing interests. 780

781




22

Figures and Tables 782 783 Table 1. Pseudovirus neutralization by 2H10, 2H10-F, 2H10-RKRF and bi-2H10. IC50s are 784

indicated in µg/ml. 785

786

Tier 2H10 wt 2H10-F 2H10-RKRF Bi-2H10 2F5 VRC01

NL4-3 1 25.20 18.68 9.15 1.84 0.16 0.20

MN-3 1 >50.00 30.38 9.36 1.39 0.03 0.06

BaL.26 1 >50.00 19.38 9.63 6.05 1.21 0.13

SF162 1a >50.00 >50.00 25.19 6.14 1.22 0.39

SF162P3 2 22.04 13.14 6.76 1.32 1.96 0.24

SC422661.8 2 >50.00 >50.00 27.93 3.79 1.00 0.27

JR-FL 2 44.65 16.93 6.95 1.49 0.97 0.11

JR-CSF 2 >50.00 21.66 10.85 2.85 1.24 0.37

QH0692.42 2 >50.00 >50.00 >50.00 >50.00 1.20 1.21

THRO4156.18 2 >50.00 >50.00 >50.00 >50.00 >50.00 3.84

787

788




23

789

790

Fig. 1. Crystal structure of gp41FP-TM in complex with 2H10. 791

A, Ribbon presentation of gp41TM-FP in complex with 2H10. Color-coding of the different 792

segments is as indicated in the gp41 scheme (Fig. S1A), the 2H10 nanobody is colored in green. 793

B, Close-up of the interaction of gp41FP-TM with 2H10. Residues in close enough contact to make 794

polar interactions are shown as sticks. 795

C, D, E, Ribbon diagram of the individual protomers named chain A, B and C. Residues within the 796

FPPR and MPER hinge regions are indicated. 797

798




24

799 800

Fig. 2. Vesicle-vesicle fusion inhibition by 2H10, bi-2H10 and 2F5. 801

(A) Time course of the lipid-mixing assay using fusion-committed vesicles. At time 30 sec (‘+N-802

MPER’), peptide (4 µM) was added to a stirring solution of unlabeled vesicles (90 µM lipid), and, 803




25

after 120 sec (indicated by the arrow), the mixture was supplemented with N-NBD-PE/N-Rh-PE-804

labeled vesicles (10 µM lipid). The increase in NBD fluorescence over time follows the dilution of 805

the probes upon mixing of lipids of target and primed vesicles (+N-MPER trace). NBD increase 806

was substantially diminished in samples incubated with bi-2H10 (2 µg/ml) prior to the addition of 807

the target vesicles (+bi-2H10, dotted trace), and totally absent if unlabeled vesicles were devoid of 808

peptide (‘no peptide’ trace). 809

(B) Left: Kinetic traces of N-MPER-induced lipid-mixing comparing the blocking effects of 2H10-810

RKRF, bi-2H10 and Fab 2F5. 811

(C) Absence of effects on lipid-mixing when vesicles were primed for fusion with the C-MPER 812

peptide, devoid of 2H10 and 2F5 epitope sequences. Antibody concentrations were 20 µg/ml in 813

these assays. 814

(D) Dose-response plots comparing the inhibitory capacities of 2H10-RKRF and bi-2H10 (purple 815

and green traces, respectively). Levels of lipid-mixing 20 or 300 sec after target vesicle injection 816

were measured (initial rates D and final extents, E) and percentages of inhibition calculated as a 817

function of the Ab concentration. The dotted line and empty symbols correspond to the effect of bi-818

2H10 when the concentration of the component 2H10 was plotted. The slashed vertical lines mark 819

the 2H10-to-peptide ratios of 1:6 and 1:3. Plotted values are means±SD of three independent 820

experiments. 821

822




26

823 Fig. 3. Gp41FP-TM interaction with bnAbs LN01 and 10E8 824

A, Cα superposition of the MPER peptide structure in complex with LN01 (pdb 6snd) onto chain 825

C-C of gp41FP-TM-2H10 structure. The lower panel shows a close-up of the interaction oriented 826

with respect to gp41 F673. 827




27

B, Cα superposition of the MPER peptide structure in complex with 10E8 (pdb 5iq7) onto the 828

corresponding chain C-C of gp41FP-TM. The lower panel shows a close-up of the interaction in 829

the same orientation as in A. 830

C, Bio-layer interferometry (BLI) binding of gp41FP-TM to 10E8 and D, to LN01. 10E8 binding 831

was fit to 1:1 model and for LN01 a steady state model was employed for fitting the data. For 10E8 832

binding, gp41FP-TM was used at concentrations from 0.2 to 25,6 nM and for LN01 binding 833

gp41FP-TM concentrations ranged from 39 to 625 nM. 834

835




28

836

837 Fig. 4. Interactions within the final post fusion conformation of gp41FP-TM modeled by MD. 838

A, Model of gp41FP-TM (Fig. S7C) after 1µs MD simulation in a bilayer. Phosphate groups of the 839

phospholipids are shown as orange spheres to delineate the membrane boundaries. 840

B, Close up on the MPER and FPPR flexible regions. 841

C, Close-up of the interaction of FP (residues 514-524) and TM (residues 681-692) viewed along 842

the three-fold axis from the N-terminus indicating an intricate network of hydrophobic interactions 843

(left panel) and from the side (right panel). Interacting side chains are labeled and shown as sticks. 844

845




29

846 Fig. 5. Conformational transitions of gp41 that lead to membrane apposition and membrane 847

fusion. 848

A, Representation of the different domains of gp41 with the residue numbers delimiting each 849

domain as indicated. The same color code has been used in all the figures. 850

B, Ribbon presentation of the Env prefusion conformation (pdb 5fuu), gp41 is constrained by gp120 851

in its native conformation. The structure of native gp41 lacks the MPER and TM regions. MPER is 852

spanning a distance of 1.5 nm 98 . 853




30

C, Ribbon of native gp41, one chain is colored according to the sheme in A and the other two chains 854

are shown in grey. 855

D, Binding to cellular receptors CD4 and subsequently to CXCR4/CCR5 induces a series of 856

conformational changes that eventually leads to the dissociation of gp120. During this process, 857

HR1, FPPR and FP will form a long triple stranded coiled coil extending 11 nm towards the target 858

cell membrane. In a first step HR2 may keep its prefusion conformation in analogy to a similar 859

intermediate, activeted influenza virus HA structure 45. Alternatively, HR2 may dissociate and form 860

a more extended conformation in agreement with locked gp41 structures bridging viral and cellular 861

membranes that bridge distances of 11 to 15 nm 46. 862

E, Bending of HR1 and HR2 will result in the six-helical bundle core structure bringing cellular 863

and viral membranes into close apposition with the 3 FPs anchored in the cellular membarne and 864

the 3 TMs anchored in the viral membrane, the gp41 conformatio represented by the gp41FP-TM 865

structure. This intermedaite gp41 conformation may have brought both membranes into close 866

apposition or may have already induced hemifuison as indicated in F. 867

G, Further reolding of FPPR-FP and MPER-TM results in the final extremely stable post fusion 868

conformation. This thus suggests that rearrangment of the membrane anchors plays crucial roles in 869

lipid mixing, breaking the hemifusion diaphragm to allow fusion pore opening. Boundaries of the 870

lipid layers are shown with orange sphere representing the phosphate atomes of the lipids present 871

in the MD simulation (snapshop taken after 1µs MD simuation). 872




ReferencesAuthor Contributions: W.W. conceived and designed the study. J.L.N. supervised the lipid mixing experiments performed by J.T. A.T. supervised neutralization assays performed by N.F. C.C. and F.D. performed molecular dynamics simulation experiments. D....

available under acc-by-nc-nd 4.0 international license · 2020. 11. 12. · 2 biofisika institute...

Documents